A multiresolution approach to automated classification of protein subcellular location images

Amina Chebira, Yann Barbotin, Charles Jackson, Thomas Merryman, Gowri Srinivasa, Robert F. Murphy, Jelena Kovacevic

Research output: Contribution to journalArticle

Abstract

Background: Fluorescence microscopy is widely used to determine the subcellular location of proteins. Efforts to determine location on a proteome-wide basis create a need for automated methods to analyze the resulting images. Over the past ten years, the feasibility of using machine learning methods to recognize all major subcellular location patterns has been convincingly demonstrated, using diverse feature sets and classifiers. On a well-studied data set of 2D HeLa single-cell images, the best performance to date, 91.5%, was obtained by including a set of multiresolution features. This demonstrates the value of multiresolution approaches to this important problem. Results: We report here a novel approach for the classification of subcellular location patterns by classifying in multiresolution subspaces. Our system is able to work with any feature set and any classifier. It consists of multiresolution (MR) decomposition, followed by feature computation and classification in each MR subspace, yielding local decisions that are then combined into a global decision. With 26 texture features alone and a neural network classifier, we obtained an increase in accuracy on the 2D HeLa data set to 95.3%. Conclusion: We demonstrate that the space-frequency localized information in the multiresolution subspaces adds significantly to the discriminative power of the system. Moreover, we show that a vastly reduced set of features is sufficient, consisting of our novel modified Haralick texture features. Our proposed system is general, allowing for any combinations of sets of features and any combination of classifiers.

Original languageEnglish (US)
Article number210
JournalBMC Bioinformatics
Volume8
DOIs
StatePublished - Jun 19 2007

Fingerprint

Multiresolution
Classifiers
Proteins
Protein
Proteome
HeLa Cells
Fluorescence Microscopy
Classifier
Textures
Texture Feature
Subspace
Fluorescence microscopy
Learning systems
Neural networks
Decomposition
Datasets
Demonstrate
Machine Learning
Neural Networks
Sufficient

ASJC Scopus subject areas

  • Structural Biology
  • Biochemistry
  • Molecular Biology
  • Computer Science Applications
  • Applied Mathematics

Cite this

A multiresolution approach to automated classification of protein subcellular location images. / Chebira, Amina; Barbotin, Yann; Jackson, Charles; Merryman, Thomas; Srinivasa, Gowri; Murphy, Robert F.; Kovacevic, Jelena.

In: BMC Bioinformatics, Vol. 8, 210, 19.06.2007.

Research output: Contribution to journalArticle

Chebira A, Barbotin Y, Jackson C, Merryman T, Srinivasa G, Murphy RF et al. A multiresolution approach to automated classification of protein subcellular location images. BMC Bioinformatics. 2007 Jun 19;8. 210. https://doi.org/10.1186/1471-2105-8-210
Chebira, Amina ; Barbotin, Yann ; Jackson, Charles ; Merryman, Thomas ; Srinivasa, Gowri ; Murphy, Robert F. ; Kovacevic, Jelena. / A multiresolution approach to automated classification of protein subcellular location images. In: BMC Bioinformatics. 2007 ; Vol. 8.
@article{fb904c0eb5e34a87a2a9d721f69b0034,
title = "A multiresolution approach to automated classification of protein subcellular location images",
abstract = "Background: Fluorescence microscopy is widely used to determine the subcellular location of proteins. Efforts to determine location on a proteome-wide basis create a need for automated methods to analyze the resulting images. Over the past ten years, the feasibility of using machine learning methods to recognize all major subcellular location patterns has been convincingly demonstrated, using diverse feature sets and classifiers. On a well-studied data set of 2D HeLa single-cell images, the best performance to date, 91.5{\%}, was obtained by including a set of multiresolution features. This demonstrates the value of multiresolution approaches to this important problem. Results: We report here a novel approach for the classification of subcellular location patterns by classifying in multiresolution subspaces. Our system is able to work with any feature set and any classifier. It consists of multiresolution (MR) decomposition, followed by feature computation and classification in each MR subspace, yielding local decisions that are then combined into a global decision. With 26 texture features alone and a neural network classifier, we obtained an increase in accuracy on the 2D HeLa data set to 95.3{\%}. Conclusion: We demonstrate that the space-frequency localized information in the multiresolution subspaces adds significantly to the discriminative power of the system. Moreover, we show that a vastly reduced set of features is sufficient, consisting of our novel modified Haralick texture features. Our proposed system is general, allowing for any combinations of sets of features and any combination of classifiers.",
author = "Amina Chebira and Yann Barbotin and Charles Jackson and Thomas Merryman and Gowri Srinivasa and Murphy, {Robert F.} and Jelena Kovacevic",
year = "2007",
month = "6",
day = "19",
doi = "10.1186/1471-2105-8-210",
language = "English (US)",
volume = "8",
journal = "BMC Bioinformatics",
issn = "1471-2105",
publisher = "BioMed Central",

}

TY - JOUR

T1 - A multiresolution approach to automated classification of protein subcellular location images

AU - Chebira, Amina

AU - Barbotin, Yann

AU - Jackson, Charles

AU - Merryman, Thomas

AU - Srinivasa, Gowri

AU - Murphy, Robert F.

AU - Kovacevic, Jelena

PY - 2007/6/19

Y1 - 2007/6/19

N2 - Background: Fluorescence microscopy is widely used to determine the subcellular location of proteins. Efforts to determine location on a proteome-wide basis create a need for automated methods to analyze the resulting images. Over the past ten years, the feasibility of using machine learning methods to recognize all major subcellular location patterns has been convincingly demonstrated, using diverse feature sets and classifiers. On a well-studied data set of 2D HeLa single-cell images, the best performance to date, 91.5%, was obtained by including a set of multiresolution features. This demonstrates the value of multiresolution approaches to this important problem. Results: We report here a novel approach for the classification of subcellular location patterns by classifying in multiresolution subspaces. Our system is able to work with any feature set and any classifier. It consists of multiresolution (MR) decomposition, followed by feature computation and classification in each MR subspace, yielding local decisions that are then combined into a global decision. With 26 texture features alone and a neural network classifier, we obtained an increase in accuracy on the 2D HeLa data set to 95.3%. Conclusion: We demonstrate that the space-frequency localized information in the multiresolution subspaces adds significantly to the discriminative power of the system. Moreover, we show that a vastly reduced set of features is sufficient, consisting of our novel modified Haralick texture features. Our proposed system is general, allowing for any combinations of sets of features and any combination of classifiers.

AB - Background: Fluorescence microscopy is widely used to determine the subcellular location of proteins. Efforts to determine location on a proteome-wide basis create a need for automated methods to analyze the resulting images. Over the past ten years, the feasibility of using machine learning methods to recognize all major subcellular location patterns has been convincingly demonstrated, using diverse feature sets and classifiers. On a well-studied data set of 2D HeLa single-cell images, the best performance to date, 91.5%, was obtained by including a set of multiresolution features. This demonstrates the value of multiresolution approaches to this important problem. Results: We report here a novel approach for the classification of subcellular location patterns by classifying in multiresolution subspaces. Our system is able to work with any feature set and any classifier. It consists of multiresolution (MR) decomposition, followed by feature computation and classification in each MR subspace, yielding local decisions that are then combined into a global decision. With 26 texture features alone and a neural network classifier, we obtained an increase in accuracy on the 2D HeLa data set to 95.3%. Conclusion: We demonstrate that the space-frequency localized information in the multiresolution subspaces adds significantly to the discriminative power of the system. Moreover, we show that a vastly reduced set of features is sufficient, consisting of our novel modified Haralick texture features. Our proposed system is general, allowing for any combinations of sets of features and any combination of classifiers.

UR - http://www.scopus.com/inward/record.url?scp=34547650901&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547650901&partnerID=8YFLogxK

U2 - 10.1186/1471-2105-8-210

DO - 10.1186/1471-2105-8-210

M3 - Article

C2 - 17578580

AN - SCOPUS:34547650901

VL - 8

JO - BMC Bioinformatics

JF - BMC Bioinformatics

SN - 1471-2105

M1 - 210

ER -